Thinking Critically about Racial Identity Data in Perinatal Health Research: An Example with Loss of Control Eating and Tobacco Dependence

Purpose: Racial inequities in perinatal health necessitate additional research. However, limited guidance exists on how to group racial identity data for analyses. We explored methods for grouping racial identity in two prior perinatal studies (LEAP and STARTS) that examined modifiable contributors to cardiovascular disease: loss of control (LOC) eating and nicotine dependence. Methods: Participants were pregnant individuals in two studies (LEAP: N=257; STARTS: N=300). LOC was assessed throughout pregnancy using the Eating Disorder Examination-Pregnancy Version. Nicotine dependence was assessed with the Fagerstrom Test for Nicotine Dependence (FTND) in late pregnancy. Participants self-reported racial identity using NIH-defined categories at baseline. We explored five methods of grouping participants by racial identity. The relation of racial identity to LOC eating or FTND score was examined in separate regression models for each method of grouping racial identity. Results: Black-identifying participants, whether Black-only (OR=1.88, p=.024) or Black and additional racial identities (OR=2.09, p=.006), were more likely to experience LOC eating than individuals identifying as White-only. However, when Black-only participants were compared to those identifying as White and additional racial identities, this effect was attenuated (OR=1.55, p=0.101). Findings were less susceptible to changes based on racial identity groupings when examining associations between racial identity and FTND. Conclusions: Among pregnant individuals, the magnitude of the associations between racial identity and LOC and, to a lesser extent, nicotine dependence, varied by racial identity grouping method, demonstrating the importance of careful considerations when examining racial identity data and highlighting the need to examine systemic factors underlying inequities.